摘要
针对无人机在复杂多障碍环境下的路径规划问题,提出基于改进粒子群的优化算法。首先,通过统一障碍物环境建模、优化适应度函数,并采用混沌粒子初始化使粒子群多样化,增强了算法的稳定性;然后,用自适应加速度系数替代传统粒子群算法的加速度常数,避免陷入局部极小值,同时提高了算法收敛到全局最优解的效率;最后,用无人机运动编码代替传统粒子群算法中粒子搜索轨迹的编码方式,提高解的最优性,搜索最优解路径。仿真结果表明:当进行无人机路径规划时,改进粒子群算法可以有效解决复杂的多障碍环境中传统粒子群算法的问题,与灰狼优化算法、差分进化算法、量子粒子群算法和传统粒子群算法相比,改进后的算法在不同场景静态环境中路径寻优精度和稳定性明显提高,且与动态粒子群算法相比,新算法也能更好地适应动态环境。
For the problem of UAV path planning in complex multi-obstacle environment,an algorithm based on Improved Particle Swarm Optimization(IPSO)is proposed.Firstly,the stability of the algorithm is enhanced by unifying the obstacle environment modeling,optimizing the fitness function,and using chaotic particle initialization to diversify the particle swarm.Then,the acceleration constant of the traditional Particle Swarm Optimization(PSO)is replaced by the adaptive acceleration coefficient to avoid falling into local minimum,while improving the efficiency of the algorithm converging to the global optimal solution.Finally,the encoding method of the particle search trajectory in the traditional PSO is replaced by the encoding of UAV motion,which is used to improve the optimality of the solution and search for the optimal path solution.The simulation results show that the IPSO can effectively solve the problems of the traditional PSO in UAV path planning in the complex multi-obstacle environment.In comparison with Gray Wolf Optimization(GWO),Differential Evolution(DE),Quantum Particle Swarm Optimization(QPSO)and traditional PSO,the improved algorithm has significantly improved the path optimization accuracy and stability in different scenarios of static environments.In comparison with Dynamic Particle Swarm Optimization(DPSO),the new algorithm can also be better adapted to the dynamic environment.
作者
徐建新
孙纬
马超
XU Jianxin;SUN Wei;MA Chao(College of Aeronautical Engineering,Civil Aviation University of China,Tianjin 300000,China)
出处
《电光与控制》
CSCD
北大核心
2023年第6期15-21,106,共8页
Electronics Optics & Control
基金
中央高校基本科研业务费项目(3122019085)。
关键词
粒子群算法
混沌粒子初始化
自适应加速度系数
运动编码
路径规划
particle swarm optimization
chaotic particle initialization
adaptive acceleration coefficient
motion encoding
path planning
